RE: libpq debug log

From: "Iwata, Aya" <iwata(dot)aya(at)jp(dot)fujitsu(dot)com>
To: 'Jim Doty' <jdoty(at)pivotal(dot)io>
Cc: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>, "nagata(at)sraoss(dot)co(dot)jp" <nagata(at)sraoss(dot)co(dot)jp>, "peter(dot)eisentraut(at)2ndquadrant(dot)com" <peter(dot)eisentraut(at)2ndquadrant(dot)com>, "tgl(at)sss(dot)pgh(dot)pa(dot)us" <tgl(at)sss(dot)pgh(dot)pa(dot)us>, 'Haribabu Kommi' <kommi(dot)haribabu(at)gmail(dot)com>
Subject: RE: libpq debug log
Date: 2018-11-21 02:59:33
Message-ID: 71E660EB361DF14299875B198D4CE5423DE8007F@g01jpexmbkw25
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

Hi Jim Doty san,

Thank you for review! I'm sorry my late reply...

> Initial Pass
> ============
> + Patch applies
> + Patch builds
> + Patch behaves as described in the thread
Thank you for your check.

> When I set a path for `PGLOGDIR` that didn't exist or was not write-able,
> the patch writes no files, and does not alert the user that no files are being
> written.
I understand. I think it means that it is necessary to confirm how the setting is going well.
There is no warning method when connection string or the environment variable is wrong.

So I added following document:
+ If the setting of the file path by the connection string or the environment variable is
+ incorrect, the log file is not created in the intended location.
+ The maximum log file size you set is output to the beginning of the file, so you can check it.
And I added the process. Please see my v2 patch.

> Performance
> ===========
> I ran two permutations of make check, one with the patch applied but not
> activated, and the other with with the files being written to disk. Each
> permutation was run ten times, and the stats are below (times are in
> seconds):
> min max median mean
> not logging 50.4 57.6 53.3 53.4
> logging 58.3 77.7 65.0 65.8
Thank you for your measurement.
I'm thinking about adding a logging level so that only the necessary information can be printed by default. It was pointed out by Haribabu san's e-mail.
This minimizes the impact of logging on performance.

Aya Iwata

Attachment Content-Type Size
v2-0001-libpq-trace-log.patch application/octet-stream 22.4 KB

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2018-11-21 04:08:37 Re: [RFC] Removing "magic" oids
Previous Message David Rowley 2018-11-21 02:22:49 Re: PostgreSQL Limits and lack of documentation about them.